Method for validating cloud mask obtained from satellite measurements using ground-based sky camera

Area of Science:

• Earth and Atmospheric Sciences
• Remote Sensing
• Data Validation

Background:

• Satellite-derived Earth surface parameters are vital but susceptible to errors from cloud mask misclassifications.
• Accurate cloud masking is essential for improving the reliability of satellite products.

Purpose of the Study:

• To develop and apply a method for validating satellite-derived cloud masks using ground-based sky camera (GSC) data.
• To characterize the accuracy of cloud masks from two algorithms (MOD35 and CLAUDIA) and their impact on Moderate Resolution Imaging Spectroradiometer (MODIS) products.

Main Methods:

• Developed a GSC cloud cover algorithm using sky index and bright index.
• Validated MODIS MOD35 and CLAUDIA cloud masks against GSC cloud masks.
• Investigated error propagation effects on MODIS-derived reflectance, brightness temperature, and NDVI.

Main Results:

• MOD35 tends to classify ambiguous pixels as cloudy; CLAUDIA tends to classify them as clear.
• MOD35 cloud mask errors have a smaller impact on MODIS clear-sky products (reflectance, temperature, NDVI) than CLAUDIA.
• CLAUDIA cloud mask errors have a smaller impact on MODIS cloudy-sky products than MOD35.

Conclusions:

• Ground-based sky camera data provide a valuable method for validating satellite cloud masks.
• The choice of cloud-screening algorithm (MOD35 vs. CLAUDIA) significantly influences error propagation in downstream satellite products.
• Understanding algorithm-specific error propagation is key for selecting appropriate satellite data for specific applications.